LibreSSL PRNG Vulnerability Patched

msm1267 writes: The OpenBSD project late last night rushed out a patch for a vulnerability in the LibreSSL pseudo random number generator (PRNG). The flaw was disclosed two days ago by the founder of secure backup company Opsmate, Andrew Ayer, who said the vulnerability was a "catastrophic failure of the PRNG." OpenBSD founder Theo de Raadt and developer Bob Beck, however, countered saying that the issue is "overblown" because Ayer's test program is unrealistic. Ayer's test program, when linked to LibreSSL and made two different calls to the PRNG, returned the exact same data both times.

"It is actually only a problem with the author's contrived test program," Beck said. "While it's a real issue, it's actually a fairly minor one, because real applications don't work the way the author describes, both because the PID (process identification number) issue would be very difficult to have become a real issue in real software, and nobody writes real software with OpenSSL the way the author has set this test up in the article."

Re:Shocked I am! Shocked!

[jandrese]

That's not exactly the case, but it's close. The issue is that the SSL library has no way of knowing if the process forks other than checking the PID. If the SSL library detects a PID change, it has to reseed the RNG to avoid getting the same random values in both the parent and the child. Due to the way Unix PIDs work, you have a guarantee that the Parent and the Child will have different pids (fork() fails otherwise). However, if a grandparent forks a parent and then exits, and the parent then forks a child, there is nothing in Unix that outright prevents the child from getting the pid of the now deceased grandparent and foiling this detection so the SSL library doesn't know that a fork happened.

So it's a potential problem, but not one that likely exists in any production code. You could write test code that exploits it fairly easily by forkbombing the machine until the pid wraps before spawning the child, but in real production code it is unlikely to be an issue. Plus it was fixed.

Re:This is not how you inspire confidence

[Qzukk]

OpenSSL's RNG is used in many places separately from the SSL communication protocol itself, sometimes just for encryption in general (S/MIME) or sometimes someone just wants really random bytes.

Many servers fork twice in order to reparent to init, repeated forking is a common idiom in unixland.

Apache with MPM-prefork forks a bunch of children from a master process, which is typically itself a descendant of apachectl. In apache's case, this shouldn't be a problem since the "master-process-rng" would have recognized the fork and reinitialized on the first openssl connection, so the children are protected because they cannot have the same PID as the master-process.

Where it would be a problem would be an application or daemon that starts up, initializes the RNG, forks twice, then without this fork touching the RNG, starts forking children to do something random (say, encrypting one file per process or establishing a single SSL connection per process or something). Without having the RNG reset by the master process, one in 65534 or so processes will have the exact same RNG, because it will have inherited the original RNG untouched and be assigned the PID that created the RNG.

Re:This is not how you inspire confidence

[Bengie]

It's not a flaw in LibReSSL, it's a flaw in the portability layer that only happens in non-OpenBSD OSes with situations where the sysadmin blocks access to /dev/urandom like a tard. The only reason the portability implementation even supports this fall-back is because there are a lot of stupid sysadmins using OpenSSL and LibReSSL needs to be as much a drop-in replacement as possible. It's a flaw in the bandaid for a situation that shouldn't happen, but does.

Re:LibreSSL cannot be different by being the same

[Carnildo]

The random number generator should not be seeded only by a PID.

The PID is used as an absolute last-ditch fallback in the case that no other sources of randomness are available. In order for this to happen, /dev/urandom needs to be inaccessible, the KERN_RANDOM sysctl needs to be unavailable, gettimeofday() needs to fail, and clock_gettime() needs to fail.

If you're running on a system that crippled, you've really only got two choices: try seeding using the PID, or use an unseeded RNG. Or follow Theo's advice and get yourself a real operating system.